If you’ve found this blog, chances are you are looking for some answers. You have questions. Lots of questions! Like… What is GA4? What is the fuss all about? What does it do? How should I approach it? What do I need to do?
Well dear reader, this post aims to answer some of those questions.
You probably saw the announcement from Google about this shift to a ‘new analytics’ in October 2020. You have also, no doubt, noticed the little arrow in your property settings calling for you to ‘upgrade’. You may have even clicked that button, but it opened more threads of questions than you thought.
This post aims to give you a short summary of how GA4 came to existence, using the analogy of methods of transportation (it will make sense I promise). There are some really cool opportunities with GA4, so I am going to share the top highlights for me so far.
We will talk about the key differences between GA Universal Analytics and GA4, this will help to set your expectations and help to navigate the new version of GA. And, if you stick with me until the end, we will finish up with some steps on how to approach moving from Universal Analytics to GA4.
Humour me, and think of analytics as a method of transportation. This analogy is going to help explain how we got to where we are today, and show the work we have ahead of us.
If analytics was a mode of transportation, then Urchin, when it delighted us back in 2005 would have been a bike. We were happy enough, we could get around, but it was hard work and if the weather was bad you got soaked to the bone.
Then around 2007 Google said to us, be gone with your bike, we have something faster, and we got an upgrade, a scooter. We were all delighted, scooting around our website data was faster, and it had an engine! But that only got us so far. We wanted more.
Google listened, and around 2012, GA got an upgrade and Google presented us with the keys to our analytics car. This is our Universal Analytics. We all know how the car works, either from just sitting as a passenger or as the driver. Some of us have a better serviced car than others, some have a sports car (hi GA360), either way, we are all familiar with the car.
Then, October 2020 rolls around, and GA announces GA4. The 4th version of their analytics programme, and, despite being called an ‘upgrade’ it is a helicopter!
You see, whilst UA and GA4 share similarities, think cars/ helicopter, they both have doors,seats, seatbelts, a dashboard and whatnot, but, GA4 runs on a totally different model to UA.
Hard truth, our UA car is going to be 10 years old next year, which in tech land, is rather old. For a number of reasons, this 4th version was overdue. Google Search Console and Google Ads have all had quite significant changes over the years, and now it’s GA’s turn.
Google is going all in on investing into their GA4 helicopter, which means no more investment on our UA car. This means we are going to slowly start to notice the car being a bit jumpy, and slow to start, it won’t be serviced anymore, it is going to start to break down. We all need to skill up, we all need to learn how to fly a helicopter.
Oh, my, new features.
Obviously measuring your marketing and website is important,and adopting GA4 early will give an advantage. But let’s face it, no one likes change. Change is hard. If you are in the research phase of GA4 and are feeling a little overwhelmed, that is normal, find some comfort in knowing that everyone, and I mean everyone, is going to have to make the switch at some point.
Whilst change is hard, change also brings new opportunities.
On that note, let me share some things to get excited about, which I hope answered the question “what is all the fuss about”.
Let’s dive in 🙂
One main advantage of GA4 is that all users of the product will have access to a BigQuery streaming export. This differs from Google Analytics today, where only Google Analytics 360 customers are able to view hit-level data via BigQuery.
If you are new to BigQuery, the best way to describe it is to think of BigQuery as the black box to your helicopter. It is going to store all your rawdata, everything. BigQuery will become your new best friend for capturing the data you are interested in, but also for joining additional data that you were not able to send with your implementation.
It is well worth setting up your BigQuery and GA4 account as soon as you can as there are no billing charges associated with exporting data from GA4 to BigQuery. They let you export a free instance to a BigQuery sandbox, if you exceed the sandbox limits, you will need to pay for the usage (queries and data storage) the cost for doing this for most sites is minimal. You can find out more about BigQuery here, and if you are worried about BigQuery costs you can control them.
My personal take on this is that it is a really good thing. At some point, everyone I have worked with gets to a point where they need to get data out of GA to do more with it, but the barrier was always having to pay for GA 360 to get what you need.
For those of you who may not be ready to use BigQuery – yet- I would still advocate setting up and linking your GA4 account to your Big Query account as it will only start to record data in your little black box the day you set it up.
If the thought of working with BigQuery freaks you out, follow Team Simmer run by Mari and Simo Ahava and sign up for their newsletters. It has been my go-to for BigQuery learning.
Who doesn’t like a funnel?
In our Universal Analytics Car, if you have a Destination URL for a conversion goal , and the URL has a few steps that the user has to take before they reach it by way of additional website pages, then you can set up a Funnel in your Goal settings. Doing this would give you a shiny new Funnel Visualization report, to the destination Goal that you may want to create. The Funnel Visualization report is a good example that shows where people are dropping off, and you can see which pages they drop off and leave to. However, You can’t segment it. Boo.
If you set up ecommerce tracking, and followed a similar pattern (configuring your ecom settings with your conversion steps) then you got a similar report in Conversions >Ecommerce> Checkout Behavior. which you can segment, yay! However, you can’t see the pages they have left you for.
Although I always wanted more out of them they were very useful to visually see user behaviour. If you wanted to do anything more, you needed to pay for GA 360.
Now, with GA4, we get the funnels that you used to only get if you stumped up the cash for GA360- this is a massive win for us small businesses that don’t have a 6 figure analytics budget. I am excited about this because all the really good funnels were always in 360.
With GA4 you can create retroactive funnels, yes that is correct, it will apply your funnel details to your historical data.And as much as bar charts are fabulous, if you wanted to change the funnel to something that reduces the cognitive load, you can now change the visualization. Think how you would benefit from a funnel that is applied retroactively to your data, and you can look at it as a line graph to help you spot trends and changes. Erm yes please GA.
My final-wiggle-in-my-seat-excited-about -funnels came when I was playing around with them and spotted the elapsed time feature. Now, I have hated the time concept of UA for ages, it is just not that helpful the way it is calculated,basically if you go on page A and then page B it can do basic math to see how long it takes from page A to B to give you time on page. Which is fine, but what if people do not go to another page?!
GA4 have changed their concept of elapsed time, and will tell you exactly what time passed between steps (seconds, minutes,days) and you can apply this to your funnels so you will know exactly how long it takes your website visitors to complete each step.
3- Path Analysis
Did you ever look at the flow reports in UA and think, oh they look cool. But then found them really hard to work with and get any insights from? Well, GA4 has given us something called Path Analysis.
You pick an end point, purchase your goods, fill out a form, subscribe to your newsletter, whatever you want. Using this end point, GA4 will show you all the steps working backwards from that end point. You get to see all the steps/ paths your visitors did on the run up to doing the end point.
This for me is brilliant, I have tried in the past to work with flow reports, but they were too rigid, and normally sampled to hell. I then tried to build sequential segments and use data to work out the steps. This takes out the guesswork, and you can add segments to this report.
Think about the use cases for this? Blending funnels with elapsed time you can see how long it takes people to do the thing you want them to do. Take that end point and see all the steps that lead them to the final point. Then add a segment to see how different cohorts behave. Ah-mazing!
4- 30 Goals (and they are flexible!)
With UA, we had 20 goals, and if you have ever built a goal, you will share the same frustrations as I had when you realized you can’t delete goals, you could only edit them.
GA4 gives you 30 goals, and trust me, you can use them all up, and will want to use them all up when you see what you can build in GA4.
Firstly, the Goals in GA4 can be toggled on or off, and you can archive them if you do not need them anymore. The best part for me is the sheer flexibility and potential of these conversions.
One example of this flexibility is the ability to create a conversion from a segment, or from a blend of events that you have been collecting.
Let’s say you are an ecommerce site, and you have a goal for ‘bought the product’.
Well, you can go deeper than that now, you can use a Recency Frequency Value proposition model to work out your good, bad, and best customers.
If your average order value is say £50, you could work out that someone who is a really good customer has lots of visits, reads all your blog content, and spends on average £150. You can build a conversion for your big spenders. Or you want to create a goal based on a particular category or product which you are focusing on. Build that goal!
Let’s say you are a SaaS site, you could build segments to show people who are warm prospects e.g. they visit your site monthly, watch your free content and have a free trial setup. Verus your mega hot customer, finished the free trial, on the paid plan and referred a friend. Or you could build a goal based on the type of SaaS product they bought.
5- Google Signals
Most UA accounts that have Google Ads setup, usually have Google Signals enabled. GA4 has the same sync, but it is using it in a slightly different way to UA. Google signals is Google’s identity graph, and it takes session data from websites and apps that Google associated with users who have signed into and opted in/ turned on Ads Personalisation.
Used with GA4 in addition to powering your cross device reports, it will be used to help fill in more information about your website visitors. This can be used to power your remarketing and ad personalisation.
6 – Remarketing Lists now with temporary exclude rules!
When I think of remarketing, my mind wanders to the bad remarketing. You know, you go to a website for a hot second, and they follow you around the web.
It could be so much better, and GA4 is giving you more flexibility in how you approach remarketing.
I am all for opted-in, well put together , relevant remarketing. There is a new feature in GA4 where you can build an audience and define a rule to temporarily exclude users for a set amount of time.
For example, being a coffee addict, I go to a website and buy a bag of coffee beans to use at home. Typical use of the coffee beans would see me wanting to buy a new bag in 4 weeks time. That website could build an audience that is set to exclude me from seeing their remarketing for a month, and then when that time is up, I am eligible to see the ads again.
Key Differences between UA and GA4
Now we know where we are in terms of the analytics journey, using my analogy of modes of transportation. We covered some of the new features, and I hope you got excited about some of the things we can do in GA4, maybe even sparked some ideas on how to use it for your business. Now, let’s talk about the key differences between our UA car and GA4 helicopter.
Data Model Universal Analytics
Our data model in UA is a hit based model and runs on the basis of a user (someone who has visited your website), and their sessions (how many visits your user makes to your site). Say you have one user who visits your website three times over a period of time – this would be counted as three sessions. When a user pops up on your site and their visit is recorded as one or more sessions, anything they interact with will trigger something to fire in the code, and this will be recorded as a hit, an interaction. For example, when a specific page was loaded, video played, pdf downloaded etc.
GA4 is the result of Firebase, Google’s analytics product for tracking apps. Web+ App which was rebranded to GA4, adopted the firebase model which was all event based. Therefore Google is moving away from the hit based data model and moving towards this User/Event data model.
Event structure UA
In Universal Analytics we have a data model that is built on User’s, Session’s and Hit’s, we also have Events. The structure for Events which you would find in your Behaviour reports follow the rules for Category, Action and Label. The Category is your broad bucket eg Video. The Action describes the ‘doing’ i.e what is the action that we are tracking eg played video, and the Label helps to provide context to the ‘thing they just did’ eg the name of the video that was just played.
As our data model for GA4 is all User and Event based, everything is an event. Like, everything. This for me is going to take some rewiring/ recalibrating, as I have used GA UA for sooooo long, it is a shift for my brain to go pageview (which was a hit) to pageview (now an event). Although, like most things, when you start working on it, you do get used to it.
We now have Event Names and Event Parameters. This is a big difference as the Category, Action, Label hierarchy has gone, and with an event based data model, everything is captured as an event.
Your account structure in UA as you know is built on Account, Properties, and Views. The account structure in GA4 is Account, Property and within your Property you can define a Data Stream, which you can think of as an equivalent for a reporting view.
The reporting interface is completely new as well. Remember, we are moving from our Universal Analytics ‘car’ where our reports looked like Audience, Acquisition, Behaviour, Conversion. Well, you ain’t driving a car, you are flying a helicopter now.
The User Interface for GA4 is different, don’t expect it to be a clone of UA.
You have a number of Standard Reports in Lifecycle.
- Acquisition = how did your visitors find your site?
- Engagement = what did they do on your website?
- Monetization= did they make you any money?
- Retention= do they come back?
There are User reports under this that show Demographics and Tech data (similar to what we would get in UA Audience Reports). There is an Event section that will show all the Conversions you have set up (remember you get 30 now!) and a list of all the Events you have running on your GA4 collection and configuration.
Then you will notice a section called Explore, this holds the Analysis Hub. This is where you can use all the event data you are collecting to build those funnels and path analysis reports that we went through at the start of this post.
What do you do now?
Bottom line, we are all going to need to skill up. You can’t just walk up to a helicopter, sit in the seat and fly to the shops. You are going to need to get a licence to fly. As with any licence you have a blend of theory and practice.
Theory : we are all going to have to learn how this data model works, and how to plan strategically for your next measurement strategy phase.
Practical: once you know how it works and what you need to do, time to roll up your sleeves and start to ‘do it’, make the properties, edit the configuration, build the events, create the conversions.
How and when you do this, well that is up to you. This change is going to happen, and change is hard.
So, work at your own pace, and my advice is to work in phases.
I am working on a roadmap that sets out a number of Theory and Practical items in set phases. This is a quick one liner on each phase.
Phase 1 is around setting up your Dual Tagging. Doing the basic configuration so you are collecting data and have some historical data to work with.
Phase 2 would focus on customising your setup, adding more events to fit your business needs.
Phase 3 is to start to use the reports in GA4 to review and compare between your UA reports
Phase 4 you can now start to enhance your setup, build your temp exclusion lists, build your segments and insights.
Phase 5 is when you are ready to dip your toe in the black box of data, BigQuery.
I have been encouraging my students and clients to start to focus on Phase 1. Learn the basics and get the core configuration done so you have something recording data in the background, even if it is just the automatically collected events that you get when you set up your GA4 config tag. And hey, you have already made a start by finding and reading this post.
You may also want to start to edge in a few key metrics in your current reports to help with the shift. For example, your current website data reports, maybe add in the User metric instead of focusing just on Sessions?
Start to look at the Demo Account. This is a sandbox account, where you can see the Google Merchandise GA4 setup.
One thing is for sure, we are all going on a journey.
If you found this post useful and want to hear about GA4 content from us, then sign up to our newsletter below.